并行回火用于基于扩散的组合优化
Parallel Tempering for Diffusion-Based Combinatorial Optimization
- ETAS Research（ETAS研究院）
- University of Stuttgart（斯图加特大学）
- Max Planck Research School for Intelligent Systems (IMPRS-IS)（马克斯·普朗克智能系统研究学院）
- School of AI and Advanced Computing, Xi’an Jiaotong-Liverpool University（西安交通大学利物浦大学人工智能与先进计算学院）
- NEC Laboratories Europe（NEC欧洲实验室）
机构由 AI 辅助整理，请以论文原文为准。
AI总结：
本文提出PT-Denoise，一种通过并行回火使去噪轨迹交互的推理时方法，在不重新训练的情况下动态重分配温度，提升图组合优化问题求解质量，且计算开销极小。
AI中文摘要：
离散扩散模型已成为一种强大的范式，通过学习采样高质量解来解决图上的组合优化（CO）问题。一种常见的推理时方法是独立生成多个候选解并返回性能最佳的样本，这以增加计算成本为代价提高了解的质量。在本工作中，我们引入了PT-Denoise，一种推理时程序，它允许这些并发的去噪轨迹通过并行回火进行交互，而无需对底层去噪器进行重新训练或微调。我们的方法为每个扩散过程分配一个温度，并允许过程根据其相对性能交换温度。这动态地将有希望的、低能量的轨迹重新分配到更冷、更集中的采样区域，同时允许高能量状态通过随机探索逃离局部最小值。在典型图结构CO问题上的实验表明，我们的方法持续提高了找到的最佳解的质量，同时仅增加了最小的计算开销。
英文摘要：
Discrete diffusion models have emerged as a powerful paradigm for solving combinatorial optimization (CO) problems on graphs by learning to sample high-quality solutions. A common inference-time approach is to generate multiple candidate solutions independently and return the best-performing sample, improving solution quality at the expense of an increase in computational cost. In this work, we introduce PT-Denoise, an inference-time procedure that allows these concurrent denoising trajectories to interact through parallel tempering, without requiring retraining or fine-tuning of the underlying denoiser. Our method assigns a temperature to each diffusion process and allows processes to swap temperatures based on their relative performance. This dynamically reallocates promising, low-energy trajectories to colder, more concentrated sampling regimes while allowing higher-energy states to escape local minima through randomized exploration. Experiments on canonical graph-structured CO problems show that our approach consistently improves the quality of the best solution found, while only adding minimal computational overhead.
